bob loblaw

Some thoughts Not failing your required classes is important... but no one cares about your Graduate GPA If it's the first time they're offering this class, reasonable professors will probably dismiss a low grade as a noisy measure of your capabilities Suppose I'm wrong. Do you want to be advised by someone that is that fixated on you performing poorly in this class that your entire cohort thought was unreasonably hard? Spending time on pursuing your research interests seems to be a pretty important priority as a PhD student That said, I'm a second year PhD student as well so you might want to ask other people in your program/professors.

@manofpeace Absolutely. Having a PhD is obviously a prerequisite now for any research oriented industry role. But also given the increased competition for data science roles, I think a PhD will be valuable there as well. WRT to math, it's not all or nothing. My view is that you should at least have taken proof-based linear algebra and real analysis at a minimum. Having other classes like stochastic processes, convex optimization, measure theory, etc. will make you a more competitive candidate. Of course, the math classes you choose to take should make sense given your research interests. For example, taking an abstract algebra course may kind of seem random for some but may make sense if your interest lie in random matrices. Different programs (bio stats included) care about mathematical background to varying degrees: more theoretical a given department is (which tends to be higher ranked or whatever), the more they'll care. It also depends on how competitive your application year happens to be: UCLA Biostats, for example, weeds people out by mathematical background especially in competitive years. My program is more on the applied side so it doesn't care that much.

With all due respect, your advisor seems like the archetype of an out-of-touch academic (probably a boomer). Sorry to hear this. If I were you, I would not let him hold you hostage, ignore sunk-costs and start establishing relationships with professors you've taken courses from. It's not the end of the world to have mediocre LORs. I personally did not have great LORs but it was fine for me. Like the previous comment said, however, having a firm mathematical foundation is important (especially in a program that emphasizes probability). Also if you're uncertain about a PhD program BEFORE you apply, those thoughts are gonna be amplified once you're in the program. I'd keep that in mind.

I was in the same situation you were in. Taking courses as a non matriculated student is generally painful. I'd take a course from UIUC's NetMath! It was just what I needed and the transcript you get is identical to UIUC's (basically).

The workload varies a lot but 90 is excessive. My upper bound so far has been 60 hrs. There's also a general culture in PhD programs of over-working and doing way more than necessary IMO.

You have a great background! I would cut down on your list of schools to 7 or 8. Maybe cut some reach schools? Other notes: UCLA Biostats may be a better fit since you're more interested in applications. If you're interested in Bayesian stuff, I'd recommend UCSC as a match school. You'd most likely get in. This is completely based on "what I've heard" but UC Irvine is actually more difficult to get into than rankings suggest.

You will definitely get in somewhere for sure! What is your eventual goal? PhD? Getting a job after MS?

Your GPA in quantitative courses will matter most. Even among courses, I assume it is weighted on relevance: a B- in numerical optimization is different than a B- in probability

Hey all, I summarized my take on applying for grad school. My guidance may be more useful for “atypical” candidates or candidates whose undergrad math background is not particularly deep. I wrote most of this last year but I made recent edits to add guidance for Master's students. https://sho-kawano.github.io/2022/01/07/grad_school_guide/ Hope this is helpful to someone out there!

It highly depends on the program! ? For example, at my institution, a Master's Student isn't really expected to have an extensive background. They even cover basic probability in their first quarter. Lemme know if you have other questions!

Your MS choices look fine to me. I think your PhD program choices are total reaches. If you applied to lower ranked PhD programs you'd probably get in to one. I think that would be better personally because PhD programs provide funding

I think UC's are more competitive than the rankings suggest. If you are interested in Biostats because you're more drawn to applied research, UCSC and UCSB are also great choices. I also believe UC Davis allows you to apply to both Stats & Biostats. Since the departments share professors & the courses, I'd suggest applying to both if you really want to go to Davis. Given your mathematical background, I really think you will have somewhere to go to. My mathbackground is much inferior and I got into 3 places last year.

Thanks for commenting @BL4CKxP3NGU1N . That's reassuring since it costs so much to get one. Also @bayessays Idunno if you got the latest one but that macbook air doesn't even have a fan? So apparently they have to throttle the CPU for thermal issues.

Considering getting a M1 Macbook Pro. Besides the cost I'm concerned about potential problems with not being able to use some R packages due to compatibility issues with Apple Silicon. Anyone have any thoughts on this? Besides tidyverse I'd imagine there isn't a whole lot of packages I would need for classes... For full context, I go to a very Bayesain program so I'm guessing I'll use my CPU a fair amount.?

With your background, you may even want to apply to PhD programs. These programs will fund your studies. You can then decide later if you want to just take the masters and leave or complete your PhD (which can get you research positions in tech).

You're on a forum with those applying to PhD programs which can give you a skewed sense of your qualifications. I think you'll get into at least one of your reach schools. The fact that you went back to CC after getting a bachelors shows you care. For LORs, you've also been out of school for a while! So unless they're not looking at applications very closely (which can happen at certain schools), they'll factor that into account. GRE: My guess is that they use this to weed people out, especially for MS programs. I know it's super annoying. I personally think getting a high score is about having done a LOT problems. I recommend following this guide: https://magoosh.com/gre/90-day-gre-study-plan-math-focused. Don't take the Math Subject one. LOR: I had problems with this so I can empathize! I would say, ask one of your CC professors & ask one of your work colleagues (for MS programs this is fine). Ask the Penn State profs for one (even if it was online). Remember if they look down online education, then most university degrees granted during this pandemic era would also be worthless! Good luck!

Undergrad Institution: UC Berkeley + Community College (1 yr) Major(s): Statistics || GPA: 3.8 (overall at Berkeley) — lower for math/stats (see below) Type of Student: Domestic Asian Male GRE General Test: Q: 166 (87%) | V: 164 (94%) | W: 4.5 (80%) Letters of Recommendation: 2 teaching/adjunct faculty in math/stats + 1 Electrical Engineering Prof. Math Grades: Calculus I-II: A Multivariate Calc. + Diff Eq. + Lower Div. Linear Algebra (Community College): A Linear Algebra (after graduating ) B+ Real Analysis (via UIUC’s NetMath) : A Other Grades: Probability: A || Math Stats: A+ Statistical Computing: A+ || Linear Modeling: B+ Time Series Analysis: B+ || Statistical Learning: A+ Intro to CS: A Research Experience: A published public policy paper + Applied factor analysis research at an air pollution lab. Work Experience: 3 years of work experience in clean energy/healthcare. Miscellaneous: Received NSF Honorable Mention. Schools Applied: Only applied to Stats PhD Programs in CA. Given my interests, I would have applied to other schools with environmental/spatial/applied bayesian stats research. 1. UC Santa Cruz - Statistics / Admitted on Jan 28th / Accepted Funding: 26k/9months + 3k summer fellowship + healthcare. Health insurance included. 2Q TA & 1Q Fellowship. 2. UC Santa Barbara - Statistics & Applied Probability / Admitted on Jan 27th / Declined Funding: 22k/9months + healthcare. TA ships. 3. UC Davis - Statistics / Waitlisted, Admitted to MS with partial funding. 4. UC Riverside / Waitlisted 5. UC Irvine / Pending 6. UCLA Biostats / Rejected on Jan 27th Reflections: I came into undergrad wanting to study environmental policy. I discovered stats my sophomore year and had to go back to community college in order to switch! Even after switching to stats, I had ZERO intentions of going to grad school. I had a very un-linear path to get here. Given the rise in the competitiveness of admissions, I feel very fortunate to have acceptances into two programs in beautiful (and expensive) places. ? Advice: My guidance may be more useful for “atypical” candidates or candidates whose undergrad math background is not particularly deep. 1. Depth of math background matters a lot … Admissions are becoming even more competitive. So the depth of math background is apparently becoming more important to differentiate oneself from other applicants. The bare minimum is Real Analysis & upper div. algebra but I’d take more if time/budget allows. If you don’t have a strong math background, it’s ok. See pt 2. 2. … but there are ways to make it up, even after graduating. If you want to improve your math background, I recommend NethMath, the UIUC program run by their math department that is fully online. It is well designed for remote instruction and is cheaper than enrolling as a non-matriculated student. It allows you to take classes while working as well. Another plus is that the transcript they produce is indistinguishable from UIUC's normal classes. I had some savings so I was tutored by a Math PhD to get some guidance in proof-writing for a few months. This was SO extremely helpful for improving my mathematical maturity, though adcoms won’t care about it. If you’re interested in that, contact someone like Alexander Coward: https://edeeu.education/director/alexandercoward 3. If you need to study & make up coursework, then do it full-time Of course, not everyone has the means for this. That said, if you know you lack the math/coursework background to be a competitive applicant, then seriously consider studying full-time. Studying math while working a job is very difficult and inefficient. There’s a momentum that comes with a full-time dedication. I mostly studied part-time and half-assed my job (during this period). In hindsight, I would have just focus on studying for a few months & then find a job. Of course, the financial hit is significant but consider it good practice for living on a grad student budget. ? 4. Domestic students: apply to NSF. You’ll be forced to read journals and think seriously about your research interests. It will also make you start your application materials early. I even got to get a professor at a perspective school to write a letter of rec for me. Through this, I got to talk to them get to know them. 5. Seriously consider taking a year or two off after undergrad. A few reasons for this: I. The years in the “real world” have been so, so valuable to me. Chances are, you’ll mature a lot. II. I think some grad students struggle with choosing an advisor because they’ve never had a real boss. Working under a boss or two will give you a better idea of what "a good advisor" actually means for you. III. Learn how to adult. Make sure you live frugally. IV. You’ll have savings!

I went to Berkeley as an undergrad. It seems that Berkeley's program is designed to be two semesters long. They made the change to a shorter program recently to better serve students looking for industry jobs (~2016). In fact, MA students used to TA but no longer have time for anything but coursework. Also, I don't think a year is nearly long enough to do these things simultaneously: take PhD level courses (Berkeley PhD courses are no joke), connect with potential letter writers, and start a project. I would feel that you'd be fighting an uphill battle given that the MA website clearly states that: "(i)n extremely rare cases, a thesis option may be considered by the MA Chair". In my experience, Berkeley doesn't make anything easy (you have to test in for their BA program for christ's sake lol) so when they say "it's extremely rare" I would take their word for it. ? I would elect Stanford (longer program) or another option that enables you to connect more with potential letter writers and take more theoretical courses. Like @DanielWarlock said, having a paper is not important.

I'm always surprised how much you find out when you reach out to the departments directly. Anyway, for the UC's you mentioned: UC Riverside seems disorganized/slow. It's unclear to me if they've sent out any offers. UC Davis' target class is only around 6 this year with a record number of applicants. I'd say chances there are low. UC Santa Barbara's admissions chair is very responsive/super nice. If you're curious, I'd reach out to her.

@csheehan10 I agree with @bayessays. I'd go work a chill job save some money, cook and exercise a lot, and travel. If you want to bolster your math background: I'd recommend taking a course or two on NetMath (UIUC). You can do everything on your own schedule. It's designed to be remote and it worked really well for me. If you want to get a better letter/research: So research-heavy medical schools (UCSF, Hopkins, etc.) regularly hire for research assistants/data analysts. With a masters, you could even get a junior statistician position. That said, a lot of these junior research positions are a hit and miss: they pay poorly but can be intense. Even if your PI is nice, you're just not going to be a priority (compared to say, a PhD student). This is just my opinion but if I wanted a better letter, I would go around asking professors for unpaid work in exchange for a letter (be upfront about it), work with them for ~2 months, move on and get a normal job that pays better and chill........

@Stat Assistant Professor I went to a pure frequentist school and have almost no background in Bayes. Assuming I have a basic undergrad stats background + upper division linear algebra + real analysis.... Do you have recommendations on Bayesian Textbooks? I was looking to review the first chapters in Gelman's "Bayesian Data Analysis" (supposedly the Bayesian Bible). Also @Egnargal do you mind comparing Robert's "The Bayesian Choice" to Gelman?

Oh sorry. Should have clarified. This is for the PhD program but I assume it's the same for MS. They don't use an official waitlist function. Their website is super outdated (& frankly, horrible to navigate) so I have no idea what the typical size of the MS cohort is. All I know is that for the PhD program they got record amount of qualified candidates this year and their target class size is only 6. Goes to show ya, so much of this process is reliant on variables that are completely out of our control. ??

Can only comment as a fellow applicant but ... according to one junior faculty at Berkeley they said, it's a good, well-rounded department to consider. Unlike some of the top ranked schools, he commented that it's a good place to do applied research as well. After reviewing the faculty research interests website, I agree with his view on this. It just seems like a well-rounded, cool department (from a research perspective).

For those of you applying to UC Davis (Stats), if you haven't heard yet, it means you're unofficially waitlisted. Just FYI.

Wassup party people! Just wanted to gauge how I should view a fully funded offer from a school that includes 2 fellowships (summer & 1 quarter) in addition to TA-ing. I really wouldn't mind teaching at all; I'm looking forward to it so just wanted to gauge your thoughts. Is this pretty standard or should I be extra psyched about this?